I understand that keeping a consistent unit framework across multiple domains is complicated and not everyone will be satisfied with the choices made, like when people ask why "Radians" are not dimensionless in the Wolfram Language.
So I suppose there is a probably a reason why both
CompatibleUnitQ["Revolutions", "Radians"]
and
CompatibleUnitQ["Revolutions"/"Seconds", "Radians"/"Seconds"]
return False
, or similarly with "AngularDegrees" instead of "Radians", or a time unit different from "Seconds". It does surprise me as there are multiple domains in which this unit conversion would feel natural.
EDIT:
I think this is a case in which people would hope Mathematica could go beyond the unending academic metrology debate regarding the status of angular units and units derived from them. Wolfram Alpha recognizes the connection of revolutions with radians (and similarly with revolutions per second and radians per second):
True
in MMA 11.3.0. $\endgroup$